Hydrometallurgy 185 (2019) 133–141

Contents lists available at ScienceDirect

Hydrometallurgy
journal homepage: www.elsevier.com/locate/hydromet

Improving gold recovery from a refractory ore via Na2SO4 assisted roasting T
and alkaline Na2S leaching
Xiaoliang Liua,b, Qian Lia, , Yan Zhanga,b, Tao Jianga, Yongbin Yanga, Bin Xua, Yinghe Heb,
⁎ ⁎

a
School of Minerals Processing and Bioengineering, Central South University, Changsha, Hunan 410083, PR China
b
College of Science and Engineering, James Cook University, Townsville, Queensland 4811, Australia

ARTICLE INFO ABSTRACT

Keywords: Gold recovery from refractory gold ores with controlled roasting remained well below 80%. Na2SO4 was added
Refractory gold ore in an O2-enriched single stage roasting of a refractory gold ore to improve its gold recovery. Changes in phy-
Na2SO4 assisted roasting sicochemical properties of the calcines suggested that this reduced the sintering as well as facilitated the for-
Thermodynamic analysis mation of pores and a water soluble phase within the calcine. Thermodynamic analysis and leaching results
Alkaline Na2S leaching
demonstrated that Na2S solutions could effectively remove Sb species from the calcine. An extraction process
that includes Na2SO4 assisted roasting and alkaline Na2S leaching is shown to be able to achieve a gold recovery
of over 95% from the refractory ore.

1. Introduction
In our previous research (Liu et al., unpublished results, 2018), an
optimised O2-enriched single-stage roasting achieved the simultaneous
removal of S, As and C from a refractory gold ore, and an improved gold
recovery of 72.5%. However, it is still unsatisfactory. Analyses of cal-
cines from the controlled roasting suggested that the main causes for
the low recovery were likely to be the secondary encapsulation of gold
particles caused by the local high temperature during roasting and the
existence of Sb in the calcine.
Sintering often occurs during roasting, particularly under an O2-
enriched atmosphere due to the enormous amount of heat released from
the oxidation of S and C (Li et al., 2017). It could result in secondary
encapsulation of gold in the calcine that prevents its exposure to the
extraction medium and therefore ultimate extraction.
In refractory gold ores, Sb exists mainly as stibnite (Sb2S3) and is
difficult to remove together with S, As and C through roasting due to
the conflicting requirements (Padilla et al., 2010, 2014). In contrast
with an O2-enriched (40 vol%) roasting at 650 °C for the simultaneous
removal of S, As and C (Liu et al., unpublished results, 2018), low
oxygen concentrations (around 1–5 vol%) and high temperatures
(> 900 °C) are required to efficiently volatilize stibnite to antimony
trioxide (Sb2O3); otherwise, Sb would remain in the calcine in the form
of less volatile Sb2O3 and/or nonvolatile Sb2O4 (or SbO2) (Kyle et al.,
2012; Padilla et al., 2010, 2014; Živkovića et al., 2002). The presence of
Sb in gold ores > 0.5 wt% has been found to severely hamper or
prevent cyanidation since Sb can consume oxygen and cyanide and/or
passivate gold particles by forming oxyanions, thioanions and/or pre-
cipitates (Kyle et al., 2012; Thomas and Cole, 2016).
In order to attain an acceptable level of gold recovery, the sintering
phenomenon must be prevented during roasting, and the Sb must be re-
moved before cyanidation. Additives are often utilised to modify the
roasting process for a range of ores. Sodium sulphate (Na2SO4) has been
reported to be an effective additive in the oxidative roasting of sulphide ores
to improve the extraction of valuable metals such as Cu, Co and Ni (Guo
et al., 2009; Li et al., 2018; Prasad and Pandey, 1998; Rao et al., 2001). This
is likely attributed to the relief of sintering and/or the formation of a porous
structure for calcines during the Na2SO4 assisted roasting (Guo et al., 2009;
Li et al., 2018). The beneficial effects of Na2SO4, however, are seldom ap-
plied to the roasting of refractory gold ores. Alkaline sulphide leaching at
relatively high temperatures (> 70 °C) has proven to be suitable for re-
moving Sb from antimonial sulphide gold ores/concentrates (Awe and
Sandström, 2010; Baláž and Achimovičová, 2006; Baláž et al., 2000; Celep
et al., 2011; Gök, 2014; Li et al., 2011; Liu et al., unpublished results, 2018;
Raschman and Sminčáková, 2012; Sminčáková and Raschman, 2011;
Ubaldini et al., 2000; Zeng et al., 2018). In addition, the oxides of Sb such as
Sb2O3 can also be dissolved in alkaline solutions (King, 2005).
This paper reports our attempt to improve the gold recovery from
refractory gold ore. Na2SO4 is added to further optimise the O2-en-
riched single stage roasting. An alkaline sulphide leaching step is also
added to remove Sb. Thermodynamic analyses are used to establish
possible mechanisms for the experimentally observed improvements.

⁎
Corresponding authors.
E-mail addresses: csuliqian208018@csu.edu.cn (Q. Li), yinghe.he@jcu.edu.au (Y. He).

https://doi.org/10.1016/j.hydromet.2019.02.008
Received 3 November 2018; Received in revised form 14 January 2019; Accepted 15 February 2019
Available online 17 February 2019
0304-386X/ © 2019 Elsevier B.V. All rights reserved.
X. Liu, et al. Hydrometallurgy 185 (2019) 133–141

Table 1 temperature, pressure, and concentrations of Sb and S, the pre-

Chemical composition of the gold ore/(wt%). dominance areas of Sb species in the system can be shown on a relevant
Aua SiO2 CaO MgO Al2O3 As Fe S Sb C K2O scale in an Eh(vs. SHE)–pH diagram.
The speciation diagrams for Sb species were constructed using
18.05 27.68 4.26 5.86 2.19 1.62 19.60 20.60 1.31 6.95 2.10 Medusa-Hydra chemical equilibrium software for Windows
a
(Puigdomenech, 2004) based on an algorithm of SOLGASWATER
Unit g/t.
(Eriksson, 1979). The thermodynamic data and relevant species are
from the hydrochemical log K database of the software. Given the
system status of Eh (vs. SHE), pH, and total concentrations of Sb and S,
the concentrations (M) or mole fractions (%) of all Sb species co-ex-
isting in an equilibrium state can be estimated from their equilibrium
constants (log K).

3. Experimental

3.1. Materials and reagents

Fig. 1. XRD spectrogram of the gold ore.
2. Thermodynamic calculations
HSC Chemistry 6.0 for Windows (Roine, 2006) based on STABCAL
(Haung, 1989; Haung and Cuentas, 1989) was used to construct the
Eh–pH diagram for Sb–S–H2O system. The available thermodynamic
data are listed in Appendix A. Under a given solution condition of
The refractory gold ore was a concentrate with over 80 wt% of the
particles smaller than 74 μm in size. As shown in Table 1, it contained
Au 18.05 g/t. The weight percentages of the impurities of concern are:
As 1.62%, S 20.76%, C 6.95% and Sb 1.31%, respectively. The total
content of SiO2, CaO, MgO and Al2O3 reached ~40%, which can cause
sintering and agglomeration under high temperatures (Li et al., 2017).
The XRD spectrogram presented in Fig. 1 indicates that the main mi-
neral phases of the ore consisted of pyrite, arsenopyrite, carbon, quartz
and dolomite. No Sb phase was detected by XRD due mainly to its low
content (1.31%). A widely accepted proprietary procedure (detailed in
Section 3.3) was used to quantify the chemical phases of the ore. As
listed in Table 2, 69.18% of the As occurred as arsenopyrite, 97.57% of
the S as sulphide (mainly pyrite), and 68.2% of the C as elemental
carbon. Most (90.23%) of the Sb occurred as antimony sulphide (mainly
stibnite). In terms of the chemical phase of gold, only 1.72% of gold was
leachable while 66.87% was encapsulated in arsenopyrite, 20.44% in
sulphides (mainly pyrite), and 10.36% in oxides (mainly iron oxides).
The chemicals used in the roasting and leaching experiments in-
cluding sodium sulphate, sodium sulphide, sodium hydroxide, sul-
phuric acid and sodium cyanide were all AR grade. The purity of both

Table 2
Chemical phases of As, S, C, Sb and Au in the gold ore/(wt%).
Phase of As Elemental arsenic Arsenic oxide Arsenic sulphide Arsenate Arsenopyrite Total

Content 0.025 0.002 0.231 0.241 1.12 1.62

Distribution 1.54 0.12 14.27 14.89 69.18 100.00

Phase of S Sulphide Sulphate Elemental sulphur Total

Content 20.10 0.15 0.35 20.60

Distribution 97.57 0.73 1.70 100.00

Phase of C Carbonate Organic carbon Elemental carbon Total

Content 1.71 0.50 4.74 6.95

Distribution 24.60 7.20 68.20 100.00

Phase of Sb Antimony oxide Antimony sulphide Antimonate Total

Content 0.077 1.182 0.051 1.31

Distribution 5.88 90.23 3.89 100.00

Phase of Au Leachable goldb Encapsulated in arsenopyrite Encapsulated in sulphides Encapsulated in oxides Encapsulated in silicates Total

a
Content 0.31 12.07 3.69 1.87 0.11 18.05
Distribution 1.72 66.87 20.44 10.36 0.61 100.00
a
Unit g/t.
b
The amount of leachable gold in the ore was determined by direct cyanidation of the ore.

134
X. Liu, et al.
the nitrogen and oxygen was over 99.5%. De-ionised water was used
throughout all experiments.
3.2. Experiment methods
The experimental set-up for roasting is schematically shown in
Fig. 2. The roasting of the gold ore with a particle size of over 80 wt%
-74 μm was conducted in a quartz tube in a horizontal tube furnace.
Prior to the roasting, the temperature was elevated to 650 °C (923 K).
Nitrogen was then passed to flush the air out from the tube. The
roasting atmosphere was controlled by adjusting the volume flowrate
ratio between oxygen and nitrogen with a total flowrate of 1 L/min.
After the tube was completely flushed with O2 of 40 vol%, 50 g of the
gold ore mixed with a desired amount of Na2SO4 loaded in a porcelain
boat was pushed into the quartz tube directly below the thermocouple.
Simultaneously, a vacuum pump operated during the whole roasting
process to ensure a very slight vacuum in the tube. The exhaust gas was
scrubbed with two bottles of saturated NaOH solution. After roasting
30 min, the calcine was taken out of the tube furnace and naturally
cooled in air, and collected for further tests or chemical analyses.
Water washing, alkaline Na2S leaching, and cyanide leaching of the
concentrate or calcine were all carried out in a 1-L jacketed glass re-
actor equipped with a constant temperature bath and an overhead
mechanical stirrer (IKA EURO-STPCUS25). The water washing condi-
tions were liquid-solid mass ratio 3:1, temperature 25( ± 0.5) °C, stir-
ring speed 300 rpm, time 30 min. Before leaching with alkaline Na2S
and/or cyanide solutions, the concentrate or calcine was wet-milled in a
ball mill to be over 80 wt% -45 μm. The alkaline Na2S leaching condi-
tions were liquid-solid mass ratio 3:1, temperature 80 °C, stirring speed
300 rpm, Na2S 0–15 g/L, NaOH 20 g/L and leaching time 1 h. It should
be noted that the water washing of the calcine also occurs during the
wet grinding and/or alkaline Na2S leaching process. The cyanidation
conditions were liquid-solid mass ratio 2.5:1, temperature 25( ± 0.5)°C,
stirring speed 600 rpm, NaCN 0.3 wt%, leaching time 36 h, and pH 11,
adjusted by controlled addition of 1.0 M NaOH solution. Filtration of
the solid residues from the water washing, alkaline Na2S leaching and/
or cyanidation tests was carried out with a vacuum filter with copious
de-ionised water. The respective filter cakes were then dried in a va-
cuum oven and collected separately for subsequent tests or assays.
3.3. Analytical methods
The contents of S and C in solids were determined by a high fre-
quency IR carbon and sulphur analyser (HW2000B, Wuxi
Yingzhicheng) while the other elements in solids were assayed through
acid digestion and an atomic absorption spectrometer (AA-6800,
Shimadzu). The concentration of Au in solution was measured by ICP-
AES (PS-6, Baird). The morphology and mineralogical phases of the ore,
Fig. 2. Schematic diagram of the experimental device for oxidative roasting. 1-horizontal tube furnace; 2-quartz tube; 3-sample in a ceramic boat; 4-thermocouple; 5-
exhaust gas scrubber; 6-rotameter; 7-vacuum pump; 8-pressure regulator.
135
Hydrometallurgy 185 (2019) 133–141
calcine and residue were determined using SEM coupled with EDS
(JSM-6360LV, JEOL) and an X-ray diffractometer (D/Max 2500,
Rigaku), respectively. The size distribution of the particles was de-
termined by wet-screening using sieves of different meshes. The specific
surface area of the particles was measured with a BET surface area
analyser (Model F-Sorb 2400, Beijing Jinaipu Technical Apparatus Co.,
Ltd.). The TG/DTG/DSC curves of the gold ore were determined with a
thermogravimetry (TG/DTG) and differential scanning calorimetry
(DSC) instrument (NETZSCH STA 449C, NETZSCH Group).
The chemical phases of As, S, C, Sb and Au presented in Table 2
were determined with a proprietary but widely accepted analytical
procedure developed by Changsha Research Institute of Mining and
Metallurgy Co., Ltd. Fundamentally, the phases of elements were de-
termined from a series of treatments of cyanidation, roasting, and
leaching with NaOH and various acids including HNO3, aqua regia, HF
and HClO4. More details can be found in Xu et al. (2016).
4. Results and discussion
4.1. Reduction of sintering during roasting
4.1.1. Effect of Na2SO4 addition on physicochemical properties of calcines
Fig. 3 presents the SEM images of calcines that shows the effects of
Na2SO4 addition in roasting and water washing on the morphologies of
the calcines. When no Na2SO4 was added, Fig. 3(b) shows the obtained
calcine particles have some signs of sintering on the surface. While not
shown, the size, morphology and surface of these calcines hardly
changed from water washing, implying negligible presence of water-
soluble phase in the calcines. Comparison of Fig. 3(d) with (b), and
Fig. 3(c) with (a), suggests that the presence of Na2SO4 during roasting
caused a more severe sintering during roasting, resulting in a larger
calcine particle size as well as partial melting on the surface of the
particles. However, the phase that partially melted and sintered the
particles together is water soluble as evidenced by the smaller calcine
particle size and more porous surface seen in Fig. 3(e) and (f) after the
water wash. The differences are quantified by comparing the particle
size distribution and specific surface area between Samples S1 and S3.
The results are presented in Table 3 where Na2SO4 assisted roasting
followed by water washing have resulted in the formation of smaller
and much more porous particles with a specific surface area increase
from 2084 cm2/g to 3249 cm2/g. The EDS data for S2 and S3 listed in
Table 4 suggest that the above changes are likely caused by the for-
mation of some Na- and S-containing phases during the Na2SO4 assisted
roasting. The phases must have a low melting point to cause the visible
sintering, and must be water soluble that their quantity in the calcine S3
is much reduced after the water wash.
To determine the phases, XRD surveys were performed on the cal-
cines S1 – S3 and their spectrograms are shown in Fig. 4. Two major
X. Liu, et al. Hydrometallurgy 185 (2019) 133–141

Fig. 3. SEM images for the calcines of (a)/(b) S1, (c)/(d) S2, and (e)/(f) S3. Conditions: S1: roasting without Na2SO4 or water washing; S2: roasting with 3 wt% Na2SO4
and no water washing; S3: roasting with 3 wt% Na2SO4 and water washing.

Table 3
Particle size distribution and specific surface area of the calcines of S1 and S3.
Calcine Size fraction distribution/(wt%) Specific surface area/(cm2/g)

+74 μm 37–74 μm −37 μm Total

S1 46.64 24.78 28.58 100 2084

S3 35.92 20.81 43.27 100 3249

Table 4
EDS data for the calcines of S2 and S3/(wt%).
Spotsa Na Mg Al Si S K Ca Ti Fe O

c-1 (in S2) 7.42 1.33 2.36 5.89 14.63 0.63 1.91 0.36 36.7 28.77
e-1 (in S3) 0.8 2.59 2.75 8.36 4.82 0.87 3.48 0.46 42.17 33.68

a
Analysed spots are in correlation with those in Fig. 3.

Fig. 4. XRD spectrograms for the calcines of S1, S2 and S3. Conditions: the same
differences should be noted. Firstly, the peak for a typical sintering as those shown in Fig. 3.
phase of (Ca,Mg)Si2O6 is present in S1; its intensity and areas are much
reduced in S2 or S3. Secondly, Na2SO4 peaks appeared in S2 but not in S1
or S3.
As our previous research results established that the presence of 1988; Li et al., 2018; Lindberg et al., 2006; Shores and Fang, 1981).
(Ca,Mg)Si2O6 was a sign of sintering during roasting (Li et al., 2017). Na2S2O7 has a low melting point of 387 °C (Lindberg et al., 2006). At
Visually, Fig. 3 shows the extent of sintering is much more prominent in 667 °C, Na2S2O7 becomes unstable and decomposes to Na2SO4 and SOx
S2 than in S1. However, there is no (Ca,Mg)Si2O6 in S2. Instead, Na2SO4 again.
is present. With a melting point of 884 °C, it is unlikely for Na2SO4 to
SO2 (g) + 1/ 2O2 (g) SO3 (g) H = 98.923 kJ/ mol (1)
cause significant sintering as it is much higher than the roasting tem-
perature of 650 °C. This suggests that an intermediate low-melting Na2 SO4 (s) + SO3 (g) Na2 S2 O7 (l) H = 118.681 kJ /mol (2)
phase must have formed during roasting but eventually disappeared at
the completion of roasting. Na2 SO4 (s) + SO2 (g) + 1 / 2O2 (g) Na2 S2 O7 (l) H
Many studies suggest that Na2SO4 can react with SOx (i.e., SO2 and
= 217.604 kJ/ mol (3)
SO3) and O2 to form sodium purosulphate (Na2S2O7) as shown in Eqs.
(1)–(3) with Eq. (3) being the total reaction of Eqs. (1) and (2) (Lam, For a gold ore containing a high content of S (20.76%) under the

136
X. Liu, et al.
Fig. 5. (a) TG/DTG and (b) DSC curves of the gold ore in O2 of 40% at a heating rate of 10 °C/min.
conditions of 40% of O2 and 650 °C, besides SO2 the gas of SO3 can
readily be formed from the roasting according to Eq. (1). In the pre-
sence of Na2SO4, it is not unreasonable to expect it to react with SOx
and O2 to form Na2S2O7 (l) according to Eq. (2) or (3), resulting in a
more serious sintering. Towards the end of the roasting, the decreasing
concentration of SOx due to its continuous removal would shift the
equilibrium of Eq. (2) or (3) to the decomposition of Na2S2O7 (l) back to
Na2SO4 (s). In addition, a local high temperature higher than 667 °C
could easily occur during roasting that would lead to the decomposition
of Na2S2O7 (l) to Na2SO4 (s).
To further assess the sintering behaviour of the gold ore during
roasting, TG–DTG–DSC analysis in O2 of 40% is performed on the ore.
The results are shown in Fig. 5.
Fig. 5(a) shows that, when the temperature increased from 300 to
650 °C, the rate of mass loss of the ore in the presence of Na2SO4 was
slower than that without Na2SO4. As discussed in our previous research
(Liu et al., unpublished results, 2018), the mass loss of the ore mainly
resulted from the oxidation and removal of S, As and C from the ore.
Without Na2SO4, DTG curve has a single significant peak at ~461 °C;
with Na2SO4, an additional peak appeared at ~497 °C and the peaks
have much reduced intensities. The DSC curves from the same
Fig. 6. The schematic diagram of the mechanism of Na2SO4 assisted roasting.
137
Hydrometallurgy 185 (2019) 133–141
measurements are presented in Fig. 5(b), displaying the corresponding
heat release pattern and intensity from the roasting. Similar to the mass
loss, the addition of Na2SO4 has also changed the heat release from the
roasting reactions from a single peak at ~455 °C to two peaks at
~466 °C and ~522 °C at a reduced intensity.
Even though the exact details of the reactions are difficult to de-
termine and are not available at this stage, it is certain that the presence
of Na2SO4 has changed the course of reaction that impacted the sin-
tering and the outcome of the roasting process. Based on the result
presented in Fig. 5, a possible mechanism for Na2SO4 assisted roasting
of the gold ore is proposed and is depicted in Fig. 6. Soon after roasting
started, Na2S2O7 (l) was likely formed from the reaction of Na2SO4 with
O2 and the released SOx during roasting. Among the solid ore and
calcine particles, liquid Na2S2O7 was like a binder, causing agglom-
eration and encapsulating some unreacted ore particles. This would
reduce the gas diffusion and roasting reactions. Towards the end of the
roasting, the decreasing SOx concentration in the atmosphere because
of its continuous removal caused the formed Na2S2O7 (l) to decompose
back to Na2SO4 and release SOx, exposing the encapsulated ore particles
again for their subsequent roasting. Due to the interconversion between
Na2SO4 and Na2S2O7 (l) during roasting, the exothermic effect was
X. Liu, et al. Hydrometallurgy 185 (2019) 133–141

Table 6
Chemical phase of Sb in the calcine/(wt%).
Phase Antimony oxide Antimony Antimony Total
sulphide sulphate

Content 0.740 0.356 0.034 1.13

Distribution 65.49 31.50 3.01 100.00

Fig. 7. Gold recovery with Na2SO4 addition in roasting.

much buffed compared with the roasting without Na2SO4. This likely
resulted in the substantial reduction in the normal sintering phase, i.e.,
(Ca,Mg)Si2O6. When the roasting was complete, the sintering and ag-
glomeration of the calcine were caused mainly by the binder Na2SO4
instead of (Ca,Mg)Si2O6. With the dissolution of water-soluble binder
Na2SO4 through water washing, the sintering and agglomeration was Fig. 8. Eh–pH diagram of Sb–S–H2O system at 25 °C and 1 atm. Conditions:
[Sb] 0.031 M (calculated according to Sb content in the calcine), [S2−] 0.064 M
significantly reduced. The interchange between Na2S2O7 (l) and Na2SO4
(i.e., 5 g/L Na2S).
during roasting also resulted in a more porous calcine that benefits its
subsequent gold leaching.

4.1.2. Impact on gold recovery regions, the Sb species are soluble in the solution in the form of Sb
The gold recovery from the calcines with different additions of (OH)6− and/or SbO2−. This is confirmed in the speciation diagram
Na2SO4 during roasting is shown in Fig. 7. It can be seen that the re- (Fig. 9) that, in the Eh range of 0 V (Fig. 9(a)) to −0.25 V (Fig. 9(b)),
covery rose with an increasing amount of Na2SO4 and reached a max- the Sb species is mainly Sb(OH)6− with a small fraction of SbO2− under
imum of 83.4% at 3 wt% Na2SO4. Beyond that point, the gold recovery the alkaline conditions (pH > ~13).
remained effectively the same. This is likely that 3 wt% Na2SO4 was Sb2S3 and Sb2O3 have been found to dissolve in alkaline sulphide
sufficient to completely eliminate sintering and agglomeration during solutions according to Eqs. (4)–(6) (Awe and Sandström, 2010; Celep
roasting, and their effects on gold recovery. et al., 2011; Gök, 2014; Li et al., 2019; Raschman and Sminčáková,
2012; Ubaldini et al., 2000), which are in agreement with the above
4.2. Removal of Sb thermodynamic analysis. In addition, thermodynamic analysis also
suggests that Sb2O4 is soluble in alkaline solutions according to Eq. (7).
4.2.1. Sb in calcine and its speciation in an aqueous solution Therefore, it is expected that Sb can be removed from the calcine
As mentioned before, the existence of Sb in calcine is detrimental to through alkaline sulphide leaching.
gold recovery. Chemical analysis results on the calcine from the O2-
Sb2 S3 + 3S2 = 2SbS33 (4)
enriched single stage roasting with 3 wt% Na2SO4 show that it still has a
residual content of Sb of 1.13%, as shown in Table 5. The chemical Sb2 S3 + 6OH = SbS33 + SbO33 + 3H2 O (5)
phases of Sb are listed in Table 6, which shows that Sb existed in the
calcine mainly as oxide (65.49%), i.e., Sb2O3 and/or Sb2O4, and sul- Sb2 O3 + 2OH = 2SbO2 + H2 O (6)
phide (31.5%), i.e., Sb2S3.
Sb2 O4 + 2OH + 2H2 O = SbO2 + Sb (OH) 6 (7)
The level of Sb in the calcine is too high and needs to be removed
before the extraction for gold. A feasible approach for Sb removal from
the calcine is by selective leaching. The predominance areas of Sb 4.2.2. Gold recovery after alkaline Na2S leaching
species in an aqueous solution, based on thermodynamic analyses, are Sb removal and gold recovery via alkaline Na2S leaching of the
shown in an Eh–pH diagram in Fig. 8. It shows that, in the high pH calcine under different conditions are shown in Table 7. It shows that if
the calcine is leached in a solution containing only NaOH (20 g/L), the
Table 5 content of Sb in the residue was reduced from 1.13% to 0.52% while
Chemical composition of the calcine/(wt%). the gold extraction from the residue increased from 83.4% to 88.2%.
Aua SiO2 CaO MgO As Fe S Sb Al2O3 C K2O With the addition of only 5 g/L Na2S to the alkaline, the Sb content was
reduced to 0.15% and the gold recovery increased to 95.8%. A further
22.60 34.71 5.28 7.43 0.38 24.64 0.45 1.13 2.51 0.13 2.54 increase of Na2S would only have a slight enhancement on both the Sb
a
removal and gold recovery.
Unit g/t.
To establish the mechanism by which Sb influences the gold ex-
traction, size distribution and specific surface area of the calcine and

138
X. Liu, et al. Hydrometallurgy 185 (2019) 133–141

Fig. 9. Speciation of Sb species at 25 °C and 1 atm. Conditions: [Sb] 0.031 M, [S2−] 0.064 M; (a) Eh = 0 V, (b) Eh = −0.25 V.

Table 7
Gold extraction from the calcine treated by alkaline Na2S leaching.
Alkaline Na2S solution T/°C Time/h Residual Sb Gold recovery/
content/% %
NaOH/(g/ Na2S/(g/
L) L)

0 0 80 1 1.13 83.4
20 0 80 1 0.52 88.2
20 5 80 1 0.15 95.8
20 15 80 1 0.14 96.0

residue particles after Na2S leaching were measured. The results are
listed in Table 8. It can be seen that, although there is little change in
the particle size of the calcine, there is a significant specific surface area
increase from 6755 cm2/g to 7856 cm2/g, indicating that the calcine
became more porous after the removal of Sb. This suggests that the
increase in gold recovery is largely due to the improved exposure of
Fig. 10. Gold recoveries from concentrates and calcines.
gold particles from the Sb removal during cyanidation.
The effects of direct alkaline Na2S leaching of the concentrate, O2-
enriched single stage roasting without or with Na2SO4, and Na2S
leaching after roasting with Na2SO4 on gold recovery were also com- 95.8%. In other words, the gold recovery has progressively improved
pared, as shown in Fig. 10. It can be seen that, without any pre-treat- after each additional proposed pre-treatment.
ment, the gold recovery from direct cyanidation of the concentrate was In this investigation, a gold recovery of over 95% is achieved from a
only 12%. Alkaline Na2S leaching pre-treatment of the concentrate refractory gold ore through the addition of Na2SO4 to the optimised
improved the gold recovery to 15.6%. O2-enriched roasting with oxygen-enriched single stage roasting and an addition of an alkaline
Na2SO4, without the alkaline Na2S leaching step, then significantly Na2S leaching. While an extra processing step of alkaline Na2S leaching
increased the gold recovery to 83.4%. Addition of the alkaline Na2S will increase the production costs, this is justified from the improve-
leaching step after the roasting further increased the gold recovery to ment of gold recovery by nearly 12%. Based on this outcome, a process

Table 8
Size distribution and specific surface area of the calcine before and after alkaline Na2S leaching under the optimal conditions.
Samples Size distribution/(wt%) Specific surface area/(cm2/g)

+74 μm 45–74 μm 37–45 μm 25–37 μm −25 μm Total

Calcine 4.15 16.52 9.05 13.82 56.46 100 6755

Residue 4.55 17.34 9.43 14.23 54.45 100 7856

139
X. Liu, et al. Hydrometallurgy 185 (2019) 133–141

of the calcines suggest that, compared to roasting without Na2SO4, the

formation of sintering phase, i.e., (Ca,Mg)Si2O6 was much reduced and
the structure of washed calcine was more porous. The beneficial effect
of Na2SO4 possibly lies in the interconversion between Na2SO4 and
Na2S2O7 occurring during roasting, which not only reduces the exo-
thermic effect but also results in the formation of a water soluble phase.
Gold leaching results showed that an optimised Na2SO4 assisted
roasting could improve the gold recovery from 72.5% to 83.4%.
The presence of Sb in the calcine has been proven to be detrimental
to the extraction of gold. Eh–pH and speciation diagrams of Sb species
constructed from thermodynamic analyses suggest that the Sb oxide
and sulphide can readily dissolve in alkaline solutions. Alkaline Na2S
leaching is shown to have effectively removed the Sb and made the
calcine much more porous. This has significantly improved the ex-
posure of gold particles in the calcine and consequently the gold re-
covery to reach over 95%.
Based on the outcome of this investigation, a process flow-chart is
suggested for the processing of refractory gold ores that contain As, S, C
Fig. 11. Process flowsheet of extracting gold from refractory gold ores.
and Sb.

flowsheet shown in Fig. 11 is recommended for the processing of this Acknowledgements

type of refractory gold ores.
5. Conclusions
To further improve the gold recovery from a refractory ore, Na2SO4
is added in roasting to minimise sintering and alkaline Na2S leaching
utilised to remove Sb. Analytical results on physicochemical properties
Financial supports from the National Natural Science Foundation of
China (Grant Nos. 51574284 and 51504293), the Fundamental
Research Funds for the Central Universities of Central South University
(No. 2017zzts194) and the China Scholarship Council (Grant Nos.
201706370222 and 201606370128) are all gratefully acknowledged.

Appendix A. Free energies of formation (kcal/mol) for relevant speciesa

Species ΔG°298 (kcal/mol) Species ΔG°298 (kcal/mol) Species ΔG°298 (kcal/mol)

S 0 SO32− (a) −116.287 Sb2O5 −203.380

H2S (a) −6.607 SO42− (a) −177.907 Sb4O6 −298.624
HS− (a) 2.973 S2O32− (a) −124.825 Sb2S3 −48.626
H2SO3 (a) −128.552 S2O42− (a) −143.539 Sb2(SO4)3 −493.906
H2S2O3 (a) −128.006 S2O62− (a) −231.605 HSbO2− (a) −97.391
H2S2O4 (a) −147.385 S2O82− (a) −266.460 SbO2− (a) −82.383
HSO3− (a) −126.103 S3O62− (a) −228.814 Sb(OH)2+ (a) b −99.313
HSO4− (a) −180.524 S4O62− (a) −248.626 Sb(OH)30 (a) −154.068
HS2O3− (a) −127.183 S5O62− (a) −228.331 Sb(OH)4− (a) −194.326
HS2O4− (a) −146.862 H+ (a) 0 Sb(OH)50 (a) b −235.959
S2− (a) 20.548 OH– (a) −37.577 Sb(OH)6− (a) b −288.918
S22− (a) 19.055 O2 (a) 3.899 Sb12(OH)644− (a) b
−3085.918
S32− (a) 17.657 H2O (a) −56.678 Sb12(OH)655− (a) b
−3137.650
S42− (a) 16.589 Sb 0 Sb12(OH)666− (a) b
−3187.731
S52− (a) 15.821 Sb2O3 −150.187 Sb12(OH)677− (a) b
−3236.462
S62− (a) 15.827 Sb2O4 −190.223 Sb2S42− (a) −23.732
a
Data from the HSC Database (Roine, 2006).
b
ΔG°298 values were calculated using equation of ΔGr° = −RT ln K = ∑(νiΔGf°(i)), of which log K values are from the hydrochemical log K database of Medusa-
Hydra chemical equilibrium software (Puigdomenech, 2004).

References concentrate. Hydrometallurgy 149 (13), 23–30.

Awe, S.A., Sandström, Å., 2010. Selective leaching of arsenic and antimony from a tet-
rahedrite rich complex sulphide concentrate using alkaline sulphide solution. Miner.
Eng. 23 (15), 1227–1236.
Baláž, P., Achimovičová, M., 2006. Mechano-chemical leaching in hydrometallurgy of
complex sulphides. Hydrometallurgy 84 (1–2), 60–68.
Baláž, P., Achimovičová, M., Bastl, Z., Ohtani, T., Sánchez, M., 2000. Influence of me-
chanical activation on the alkaline leaching of enargite concentrate. Hydrometallurgy
54 (2–3), 205–216.
Celep, O., Alp, I., Deveci, H., 2011. Improved gold and silver extraction from a refractory
antimony ore by pretreatment with alkaline sulphide leach. Hydrometallurgy 105
(3), 234–239.
Eriksson, G., 1979. An algorithm for the computation of aqueous multi-component,
multiphase equilibria. Anal. Chim. Acta 112 (4), 375–383.
Gök, Ö., 2014. Catalytic production of antimonate through alkaline leaching of stibnite
Guo, X., Li, D., Park, K.H., Tian, Q., Wu, Z., 2009. Leaching behavior of metals from a
limonitic nickel laterite using a sulfation–roasting–leaching process.
Hydrometallurgy 99 (3–4), 144–150.
Haung, H.-H., 1989. Construction of Eh-pH and other stability diagrams of uranium in a
multicomponent system with a microcomputer—II. Distribution diagrams. Can.
Metall. Q. 28 (3), 235–239.
Haung, H.-H., Cuentas, L., 1989. Construction of Eh–pH and other stability diagrams of
uranium in a multicomponent system with a microcomputer—I. Domains of pre-
dominance diagrams. Can. Metall. Q. 28 (3), 225–234.
King, R.B., 2005. Encyclopedia of Inorganic Chemistry, second ed. Wiley, Hoboken, New
Jersey, United States.
Kyle, J.H., Breuer, P.L., Bunney, K.G., Pleysier, R., 2012. Review of trace toxic elements
(Pb, Cd, Hg, As, Sb, Bi, Se, Te) and their deportment in gold processing.
Hydrometallurgy 111–112 (3), 10–21.
Lam, R.K.-F., 1988. Reactions of Na2SO4 (l) and CoSO4-Na2SO4 (l) with SO3 (g) and
Al2O3(s). Massachusetts Institute of Technology, Cambridge, Massachusetts, United

140
X. Liu, et al.
States, pp. 900–1250 (K. Ph.D. Thesis).
Li, Y., Liu, Z., Li, Q., Zhao, Z., Liu, Z., Zeng, L., 2011. Removal of arsenic from Waelz zinc
oxide using a mixed NaOH–Na2S leach. Hydrometallurgy 108 (3–4), 165–170.
Li, Q., Ji, F., Xu, B., Hu, J., Yang, Y., Jiang, T., 2017. The consolidation mechanism of gold
concentrates containing sulfur and carbon at oxygen-enriched air roasting. Int. J.
Miner. Metall. Mater. 24 (4), 386–392.
Li, G., Cheng, H., Chen, S., Lu, X., Xu, Q., Lu, C., 2018. Mechanism of Na2SO4 promoting
nickel extraction from sulfide concentrates by sulfation roasting–water leaching.
Metall. Mater. Trans. B Process Metall. Mater. Process. Sci. 49 (3), 1136–1148.
Li, Y., Min, X., Ke, Y., Liu, D., Tang, C., 2019. Preparation of red mud-based geopolymer
materials from MSWI fly ash and red mud by mechanical activation. Waste Manag.
83, 202–208.
Lindberg, D., Backman, R., Chartrand, P., 2006. Thermodynamic evaluation and opti-
mization of the (Na2SO4+K2SO4+Na2S2O7+K2S2O7) system. J. Chem. Thermodyn.
38, 1568–1583.
Liu, X., Li, Q., Zhang, Y., Jiang, T., Yang, Y., Xu, B., He, Y., 2018. Simultaneous removal of
S and As from a refractory gold ore in a single stage O2-enriched roasting process.
Metall. Mater. Trans. B Process Metall. Mater. Process. Sci (unpublished results).
Liu, D., Min, X., Ke, Y., Chai, L., Liang, Y., Li, Y., Yao, L., Wang, Z., 2018. Co-treatment of
flotation waste, neutralization sludge, and arsenic-containing gypsum sludge from
copper smelting: solidification/stabilization of arsenic and heavy metals with
minimal cement clinker. Environ. Sci. Pollut. Res. 25 (8), 7600–7607.
Padilla, R., Ramírez, G., Ruiz, M.C., 2010. High-temperature volatilization mechanism of
stibnite in nitrogen-oxygen atmospheres. Metall. Mater. Trans. B Process Metall.
Mater. Process. Sci. 41 (6), 1284–1292.
Padilla, R., Aracena, A., Ruiz, M.C., 2014. Kinetics of stibnite (Sb2S3) oxidation at roasting
temperatures. J. Min. Metall. B 50 (2), 127–132.
Prasad, S., Pandey, B.D., 1998. Alternative processes for treatment of chalcopyrite–a
141
Hydrometallurgy 185 (2019) 133–141
review. Miner. Eng. 11 (8), 763–781.
Puigdomenech, I., 2004. MEDUSA: Make Equilibrium Diagrams Using Sophisticated
Algorithms. Royal Institute of Technology, Stockholm, Sweden.
Rao, K.A., Natarajan, R., Padmanabhan, N.P.H., 2001. Studies on recovery of copper,
nickel, cobalt and molybdenum values from a bulk sulphide concentrate of an Indian
uranium ore. Hydrometallurgy 62 (2), 115–124.
Raschman, P., Sminčáková, E., 2012. Kinetics of leaching of stibnite by mixed Na2S and
NaOH solutions. Hydrometallurgy 113–114 (3), 60–66.
Roine, A., 2006. Outokumpu HSC Chemistry for Windows: chemical reaction and equi-
librium software with extensive thermochemical database. In: User's Guide, Version
6.0. Outokumpu Research Oy, Pori, Finland.
Shores, D.A., Fang, W.C., 1981. Transport of oxidant in molten Na2SO4 in O2-SO2-SO3
environments. J. Electrochem. Soc. 128 (2), 346–348.
Sminčáková, E., Raschman, P., 2011. Leaching of natural stibnite using Na2S and NaOH
solutions. Int. J. Energ. Eng. 1 (2), 85–89.
Thomas, K.G., Cole, A.P., 2016. In: Adams, M.D. (Ed.), Gold Ore Processing – Project
Development and Operations. vol. 23. pp. 373–392 Roasting
developments–Especially oxygenated roasting.
Ubaldini, S., Vegliò, F., Fornari, P., Abbruzzese, C., 2000. Process flow-sheet for gold and
antimony recovery from stibnite. Hydrometallurgy 57 (3), 187–199.
Xu, B., Yang, Y., Li, Q., Jiang, T., Liu, S., Li, G., 2016. The development of an en-
vironmentally friendly leaching process of a high C, As and Sb bearing sulfide gold
concentrate. Miner. Eng. 89, 138–147.
Zeng, W., Wang, S., Free, M.L., Tang, C., Xiao, R., Liang, Y., 2018. Design and modeling of
an innovative copper electrolytic cell. J. Electrochem. Soc. 165 (14), E798–E807.
Živkovića, Ž., Štrbaca, N., Živkovića, D., Grujičića, D., Boyanovb, B., 2002. Kinetics and
mechanism of Sb2S3 oxidation process. Thermochim. Acta 383 (1–2), 137–143.